Travis Gibbs

I'm Travis, a software engineer.

Experience

  1. Gem

    Jun 2024 — Present

    Software Engineer I → Senior Software Engineer · San Francisco, CA

    Promoted L3 → L4 → L5 (senior) in under 2 years

    • Founding AI/fullstack engineer for 3 LLM-powered products, taking them from 0 → $4M+ ARR in under 2 years.
    • Owned the end-to-end retrieval and reranking stack, plus embedding pipelines across 650M profiles.
    • Fine-tuned embedding models, achieving a 10+ point recall lift over state-of-the-art baselines on internal benchmarks.
    • Led 0→1 development of a multi-agent search system enabling complex multi-step retrieval workflows.
    • Built an evaluation framework with human-aligned metrics, using in-house annotation tooling and data collection.
    • Led migration of the search product to Elasticsearch, reducing timeouts by 95% and improving response time by 5–10s.
    • Partnered with enterprise customers (Airbnb, Coinbase, Dropbox) to turn ambiguous requirements into shipped features and closed deals.
    • Co-designed an async outbox framework for reliable background task processing across distributed AI workflows.

  2. Epic Systems

    Jun 2023 — Jun 2024

    Software Engineer, Inpatient Rover · Verona, WI

    • Optimized the mobile barcode scanning UI to save 3 seconds on over 20 million monthly scans.
    • Led a project to overhaul the notification system for the mobile platform.
    • Uncovered dozens of critical bugs by building a simulated poor-wifi environment for mobile devices.

  3. Meta

    Jun 2020 — Aug 2022

    Software Engineering Intern · Remote / Boston, MA

    3 separate internships across 2020, 2021, and 2022

    • Built internal debugging and build-health tooling across the Messenger, Facebook, and Instagram teams using React, PHP, and Python; cut per-issue triage time by ~30 minutes.
    • Justified a new co-location algorithm via data tooling that cut WPS battery usage by 70%.
    • (FBU) Built Drivify, an Android app generating dynamic playlists from the Google and Spotify APIs plus GPS speed metrics.

Projects

  1. travis.show

    This site

    Static Astro site with markdown posts. Built into a Caddy container and deployed to Fly.

    • Astro
    • TypeScript
    • Docker
    • Fly.io

  2. Penelope

    Hack

    CLI proxy that screens commands from autonomous agents before they reach the shell and blocks destructive ones. Built before Claude Code had auto mode, which replaced it in my workflow.

    • Rust
    • CLI
    • Agent Safety

  3. Pumpkin

    Hack

    CLI tools for inspecting Claude Code internals, mainly reading background task output that the UI truncates. Worked well for single-terminal flows, which got harder to scale.

    • Shell
    • CLI
    • Claude Code

  4. The Elevator

    College project

    Every 64x64, 16-color image that can exist has a floor number, and the keypad takes you to it. Each pixel position runs through its own permutation table, so every code maps to one image and every image back to one code.

    • Flask
    • JavaScript
    • Generative

  5. MLB Baseball Games

    College project

    Prediction games built on data scraped from Baseball Reference and Baseball Savant. Uses audio analysis to find the moment of bat-on-ball contact in game video.

    • Python
    • Flask
    • Audio Processing
    • Web Scraping

  6. SportsCardToolLib

    College project

    Python library for scraping and querying sports card checklists, published to PyPI. Scraped 10M+ data points and exposed them through a query-builder API.

    • Python
    • REST API
    • MongoDB
    • PyPI

  7. Drivify

    College project

    Android app that builds a Spotify playlist to fit the length of your drive, using route duration and track features.

    • Android
    • Java
    • Spotify API
    • Google Maps API

Skills

Languages
  • Python
  • TypeScript
  • Kotlin
  • SQL
Infra & Data
  • Elasticsearch
  • Snowflake
  • dbt
  • Modal
  • PostgreSQL

Education

  1. Columbia University

    2019 — May 2023

    B.S. Computer Science, Fu Foundation School of Engineering (SEAS) · New York, NY

    • Concentration in AI and Intelligent Systems; minor in History.
    • First Generation Low Income (FGLI) Campus Representative.
    • Coursework: Artificial Intelligence, Natural Language Processing, Computer Vision I & II, Applied Robotics, Machine Learning, Analysis of Algorithms, User Interface Design, Intelligent and Connected Systems.

Elsewhere